Wearable Sensor Yaw Correction for Real-Time Human Pose Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motion capture systems, particularly those using 3-point tracking for VR, face limitations such as non-real-time pose tracking, lack of controller tracking, inflexibility in avatar skinning, and outdoor use restrictions, while sparse inertial systems suffer from performance issues and limited flexibility.
Innovation Solution
A method involving wearable sensors that acquire yaw measurements, calculate and correct errors using reference measurements, enabling accurate and real-time pose tracking without camera reliance and extending usability beyond outdoor environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If camera-based 3-point tracking is used for VR systems, then avatar generation is simplified, but pose tracking accuracy deteriorates and controller tracking becomes unreliable
Solution Approach 1:
The patent replaces camera-based optical tracking with an inertial measurement system using IMU sensors. The system uses accelerometers, gyroscopes, and magnetometers to directly measure body orientation and position through physical sensing, eliminating the need for camera-based visual tracking and marker detection.
Solution Approach 2:
The patent creates a virtual copy of the physical body by mapping IMU sensor data from multiple body locations to a 3D avatar model. The system reconstructs body pose by integrating inertial measurements and projecting them onto a virtual human model, achieving accurate pose tracking without direct visual observation.
2Adaptability or versatility
If sparse inertial sensors are used for motion capture, then system flexibility improves, but real-time performance deteriorates
Solution Approach 1:
The patent performs preliminary calibration by establishing the relationship between IMU sensor measurements and actual body pose during an initial setup phase. The system pre-computes transformation matrices and calibration parameters that enable real-time pose estimation without requiring complex calculations during actual motion capture operations.
Solution Approach 2:
The patent implements a feedback mechanism where the system continuously compares predicted pose from IMU integration with visual observations from cameras when available. This feedback loop corrects drift and accumulative errors in inertial integration, maintaining real-time performance while improving accuracy through iterative refinement.
3Measurement precision
If camera-based tracking is used, then controller tracking can be achieved, but outdoor use and unrestricted environment operation deteriorate
Solution Approach 1:
The patent makes the motion capture system self-sufficient by using onboard IMU sensors that do not require external infrastructure. The sensors autonomously measure body orientation and position using inertial forces, magnetic field references, and gravitational acceleration, enabling operation in any environment including outdoor settings without camera coverage.
4Measurement precision
If hand-held controllers are used for tracking, then controller position can be tracked, but the controllers must remain in camera view which restricts movement
Solution Approach 1:
The patent replaces camera-based optical tracking of controllers with direct inertial sensing. IMU sensors embedded in the controllers measure their own orientation and position through acceleration and rotation measurements, eliminating the requirement for visual line-of-sight and enabling unrestricted movement throughout the environment.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides accurate, real-time pose tracking and supports flexible avatar customization, enhancing usability across various environments and applications.
Implementation Method 1
A first measurement 140a of an orientation of the device 130 in a local magnetic frame of reference (FoR) 108a is obtained based on a magnetometer 422 of the device 130
Implementation Method 2
A second measurement 140b of the orientation of the device 130 in the spatial FoR 108b is obtained based on a gyroscope 420 of the device 130
Implementation Method 3
A second measurement 140b of the orientation of the device 130 in the spatial FoR 108b is obtained based on an accelerometer 420 of the device 130
Data Source
AI summary
A method of tracking wearable sensors attached to respective body parts of a user includes acquiring multiple yaw measurements from a wearable sensor by measurement circuitry within the wearable sensor, calculating errors in the yaw measurements based on comparisons of the yaw measurements with one or more yaw references, and correcting the yaw measurements by removing the errors.