Vehicle Head Pose Tracking Using Seat-Based 3D Position Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing head-pose tracking systems in dynamic environments, such as moving vehicles, suffer from high latency and errors due to lighting conditions and require multiple cameras, leading to inaccurate and inefficient tracking, affecting user experience.
Innovation Solution
A system and method that utilize seat position and orientation sensors to estimate a 3D volume where the user's head is likely to be, discarding outlier positions outside this volume for accurate head tracking, independent of lighting conditions, using a combination of cameras and processors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple tracking cameras are employed to track head pose, then tracking reliability is improved, but device complexity and operational cost increase significantly
Solution Approach 1:
The patent introduces an intermediary computational model that predicts head pose based on seat position and orientation data, acting as a mediator between the tracking camera and the final head pose output. This model filters and refines tracking data using biomechanical constraints, improving reliability without requiring multiple cameras. The intermediary processing layer compensates for the limitations of single-camera tracking in dynamic vehicle environments.
2Difficulty of detecting and measuring
If depth cameras are used for head pose tracking, then tracking capability is improved, but measurement precision deteriorates under varying lighting conditions such as direct sunlight
Solution Approach 1:
The patent changes the parameters used for head pose estimation by incorporating seat position and orientation data alongside tracking camera data. Instead of relying solely on depth camera measurements that are sensitive to lighting conditions, the system uses biomechanical models that predict head pose based on seat configuration, providing a lighting-independent alternative that maintains precision in direct sunlight.
3Adaptability or versatility
If existing head-pose tracking systems operate in dynamic vehicle environments, then adaptability is improved, but tracking precision and responsiveness deteriorate due to sudden movements and vibrations
Solution Approach 1:
The patent applies preliminary action by using biomechanical models to predict the range of feasible head positions based on seat position and orientation before actual head pose tracking occurs. This pre-computed constraint volume filters out impossible head positions caused by vehicle movements and vibrations, allowing the system to maintain precision in dynamic environments by anticipating valid position ranges in advance.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
Tracking means (102, 712) is utilised to estimate potential positions of a head (210, 412, 506) of a user (208, 718) sitting on a seat (204, 206, 300, 404, 504) of a vehicle (202, 406). A three-dimensional (3D) volume where the user's head is likely to be present is determined, based on a current setting of adjustable seat parameter(s), the adjustable seat parameter(s) being detected by sensor(s) (104). It is detected whether at least one of the plurality of potential positions of the head lies outside the 3D volume. When it is detected that at least one of the plurality of potential positions of the head lies outside the 3D volume, the at least one of the plurality of potential positions is considered as an outlier, and a correct position of the user's head is determined as one of the plurality of potential positions of the head that lies within the 3D volume.