Vehicle Occupant Posture Detection Using Torso-Plane Rotation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle systems struggle to accurately determine a user's posture, particularly to identify if the user is out of position, which can compromise the safety of active safety systems like airbags.
Innovation Solution
A method and apparatus that utilize sensor-based measurement data to determine a user's 3D seat-position, identify reference points on the torso, and calculate the angle of rotation of the torso-plane, employing a body posture classification algorithm to assess the user's posture, reducing computational resources compared to image comparison methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image comparison methods are used to determine user posture, then measurement precision is improved, but use of energy and computational resources increase
Solution Approach 1:
The patent segments the user's body into key reference points (shoulders, hips, head position) and processes only these critical points rather than performing full image comparison. This segmentation maintains posture detection accuracy while significantly reducing computational load and energy consumption.
Solution Approach 2:
The patent extracts essential posture information by identifying specific reference points on the user's body (shoulder positions, hip positions, head position) from sensor data. This extraction approach focuses computational resources on critical measurements rather than processing entire images, resolving the contradiction between precision and energy use.
2Measurement precision
If full image comparison is performed to determine posture, then measurement precision is improved, but productivity decreases
Solution Approach 1:
By dividing the posture detection task into identification of discrete reference points (shoulders, hips, head) rather than full image comparison, the system achieves both high precision in posture detection and improved processing speed, directly resolving the contradiction between measurement precision and productivity.
Solution Approach 2:
The patent applies partial action by performing only the necessary measurements for posture determination (key reference points) rather than complete image analysis. This partial processing approach maintains sufficient accuracy for safety applications while dramatically improving processing speed and system productivity.
3Device complexity
If sensor-based measurement data processing is simplified, then device complexity is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent extracts and processes only the essential sensor data needed for posture determination (positions of reference points on the body) rather than processing all available sensor data. This extraction strategy reduces computational complexity while maintaining the precision needed for accurate seat-position and posture determination.
Solution Approach 2:
The patent applies local quality by focusing computational processing on specific critical areas (reference points such as shoulders, hips, and head position) rather than uniformly processing all sensor data. This localized approach reduces overall device complexity while preserving measurement precision where it matters most for safety.
Data Source
Figure 1A~1B
Figure 2
Figure 3
AI summary
The invention relates to a method for determining a posture of a user (100) of a vehicle (200), comprising: Receiving sensor-based measurement data, which represent the user (100) in the vehicle (200); Determining a seat-position of the user (100), based on the measurement data; Determining positions of a plurality of reference points (P1, P2, P3, P4) of a torso (110) of the user (100) at the seat-position, based on the measurement data; Determining a torso-plane (120) of the torso (110) based on the determined positions of the plurality of reference points (P1, P2, P3, P4); Determining an angle of rotation (θ) of the torso-plane (120) around an axis (x, y, z) with respect to a reference angle of rotation; Determining a body posture by using the determined angle of rotation (θ) and a body posture classification algorithm.